Skill · Operations and quality
Recurring guest complaints
Which complaint theme is appearing more often, and which records are awaiting follow-up?
Download skill package
sikayet-tekrar-konulari/ ├─ SKILL.md ├─ assets/ │ └─ ornek-girdi.xlsx └─ ornekler/ ├─ analiz.xlsx └─ yonetim-ozeti.docx
SKILL.md instructions and sample files in one package
How to use
- Download the package.
- In an assistant that supports skills (e.g. Claude), upload the package as a skill. In other assistants, paste the SKILL.md text into the chat.
- Attach your own Excel export and send the message below. The assistant prepares the formula-based workbook and the Word summary.
Try it with the sample
Before using your own data you can try the sample input in the package. Send this message to the assistant together with the file:
Analyze the files I attached for recurring guest complaints. Which complaint theme is appearing more often, and which records are awaiting follow-up? Prepare the result as a formula-based Excel file and a one-page Word report we can use in our meeting.
Room noise rose from two cases to six. Three of the six current cases are open.
Sample files
- Sample input
ornek-girdi.xlsx - Expected Excel analysis
analiz.xlsx - Expected Word summary
yonetim-ozeti.docx
The sample files are in Turkish.
Instructions (SKILL.md)
Which complaint theme is appearing more often, and which records are awaiting follow-up?
Input and scope
Complaint list with anonymised case number, week, verified theme and final status.
Use the attached Excel/CSV files as the source. Read each one without modifying it. Identify the sheets, report scope, periods, units and required fields. Match synonymous column headers to the sample. If a missing field affects only a specific metric, still produce the other metrics. If a critical definition can be interpreted in more than one way, ask one focused question. Trace subtotals and duplicates back to the source row. Remove fields containing personal data from the output. Use the sample only when the user wants to try it with the sample data.
Analysis
- If raw records are available, deduplicate by case + theme + period. Do not count messages and status updates on the same case as new complaints. A single case can cover more than one theme. Do not present the theme total as a count of unique guests.
- Determine open/closed status from the latest record at the time of the report. Show missing statuses separately. The sample Excel starts from a count already deduplicated by theme.
- Compare periods of equal length. If the previous period is zero, give the difference in count. Without exposure data, do not derive a satisfaction rate from case counts. If exposure data is available, also show the rate per 100 stays.
Delivery and checks
On the first sheet of Analiz.xlsx, show the rows that answer the question and the calculated indicators. Keep the sources on a separate sheet with anonymised fields. Write Excel formulas for derived measures. On the Checks sheet, add the date range, column mapping, scope, missing/unmatched records and the reconciliation result. Create the Word file as Yonetim-Ozeti.docx. On one page, write the period, the main result, at most three evidenced findings and the work the responsible team should check. Reopen both files; independently check the numbers, the denominators of the ratios and the source totals. Do not report causes you cannot verify as findings. Deliver both files as downloads.
Visible fields in the Excel analysis: Theme, Previous cases, Current cases, Case difference, Open cases, Open share.
Try it with the sample
Use assets/ornek-girdi.xlsx. The expected Excel and Word examples are in the ornekler/ folder. Check specific to this sample: Room noise rose from two cases to six. Three of the six current cases are open.